Where branches are used to maintain separate lines of development, at some stage you will want to merge the changes made on one branch back into the trunk, or vice versa.
It is important to understand how branching and merging works in Subversion before you start using it, as it can become quite complex. It is highly recommended that you read the chapter Branching and Merging in the Subversion book, which gives a full description and many examples of how it is used.
The next point to note is that merging always takes place within a working copy. If you want to merge changes into a branch, you have to have a working copy for that branch checked out, and invoke the merge wizard from that working copy using → .
In general it is a good idea to perform a merge into an unmodified working copy. If you have made other changes in your WC, commit those first. If the merge does not go as you expect, you may want to revert the changes, and the Revert command will discard all changes including any you made before the merge.
There are three common use cases for merging which are handled in slightly different ways, as described below. The first page of the merge wizard asks you to select the method you need.
- Merge a range of revisions
-
This method covers the case when you have made one or more revisions to a branch (or to the trunk) and you want to port those changes across to a different branch.
What you are asking Subversion to do is this: “ Calculate the changes necessary to get [FROM] revision 1 of branch A [TO] revision 7 of branch A, and apply those changes to my working copy (of trunk or branch B). ”
If you leave the revision range empty, Subversion uses the merge-tracking features to calculate the correct revision range to use. This is known as a reintegrate or automatic merge.
- Merge two different trees
-
This is a more general case of the reintegrate method. What you are asking Subversion to do is: “ Calculate the changes necessary to get [FROM] the head revision of the trunk [TO] the head revision of the branch, and apply those changes to my working copy (of the trunk). ” The net result is that trunk now looks exactly like the branch.
If your server/repository does not support merge-tracking then this is the only way to merge a branch back to trunk. Another use case occurs when you are using vendor branches and you need to merge the changes following a new vendor drop into your trunk code. For more information read the chapter on vendor branches in the Subversion Book.
In the From: field enter the full folder URL of the branch or tag containing the changes you want to port into your working copy. You may also click to browse the repository and find the desired branch. If you have merged from this branch before, then just use the drop down list which shows a history of previously used URLs.
If you are merging from a renamed or deleted branch then you will have to go back to a revision where that branch still existed. In this case you will also need to specify that revision as a peg revision in the range of revisions being merged (see below), otherwise the merge will fail when it can't find that path at HEAD.
In the Revision range to merge field enter the list of revisions you want to merge. This can be a single revision, a list of specific revisions separated by commas, or a range of revisions separated by a dash, or any combination of these.
If you need to specify a peg revision for the merge, add the
peg revision at the end of the revisions, e.g. 5-7,10@3.
In the above example, the revisions 5,6,7 and 10 would be merged,
with 3 being the peg revision.
![[Important]](important.png)
|
Important |
|---|---|
|
There is an important difference in the way a revision range is specified with TortoiseSVN compared to the command line client. The easiest way to visualise it is to think of a fence with posts and fence panels. With the command line client you specify the changes to merge using two “fence post” revisions which specify the before and after points. With TortoiseSVN you specify the changeset to merge using “fence panels”. The reason for this becomes clear when you use the log dialog to specify revisions to merge, where each revision appears as a changeset. If you are merging revisions in chunks, the method shown in the Subversion book will have you merge 100-200 this time and 200-300 next time. With TortoiseSVN you would merge 100-200 this time and 201-300 next time. This difference has generated a lot of heat on the mailing lists. We acknowledge that there is a difference from the command line client, but we believe that for the majority of GUI users it is easier to understand the method we have implemented. |
The easiest way to select the range of revisions you need is to click on , as this will list recent changes with their log comments. If you want to merge the changes from a single revision, just select that revision. If you want to merge changes from several revisions, then select that range (using the usual Shift-modifier). Click on and the list of revision numbers to merge will be filled in for you.
If you want to merge changes back out of your working copy, to revert a change which has already been committed, select the revisions to revert and make sure the Reverse merge box is checked.
If you have already merged some changes from this branch, hopefully you will have made a note of the last revision merged in the log message when you committed the change. In that case, you can use for the Working Copy to trace that log message. Remembering that we are thinking of revisions as changesets, you should Use the revision after the end point of the last merge as the start point for this merge. For example, if you have merged revisions 37 to 39 last time, then the start point for this merge should be revision 40.
If you are using the merge tracking features of Subversion, you do not need to remember which revisions have already been merged - Subversion will record that for you. If you leave the revision range blank, all revisions which have not yet been merged will be included. Read Section 4.20.5, “Merge Tracking” to find out more.
When merge tracking is used, the log dialog will show previously merged revisions, and revisions pre-dating the common ancestor point, i.e. before the branch was copied, as greyed out. The Hide non-mergeable revisions checkbox allows you to filter out these revisions completely so you see only the revisions which can be merged.
If other people may be committing changes then be careful about using the HEAD revision. It may not refer to the revision you think it does if someone else made a commit after your last update.
If you leave the range of revisions empty or have the radio button all revisions checked, then Subversion merges all not-yet merged revisions. This is known as a reintegrate or automatic merge.
There are some conditions which apply to a reintegrate merge. Firstly, the server must support merge tracking. The working copy must be of depth infinite (no sparse checkouts), and it must not have any local modifications, switched items or items that have been updated to revisions other than HEAD. All changes to trunk made during branch development must have been merged across to the branch (or marked as having been merged). The range of revisions to merge will be calculated automatically.
Click and go to Section 4.20.3, “Merge Options”.
If you are using this method to merge a feature branch back to trunk,
you need to start the merge wizard from within a working copy of trunk.
In the From: field enter the full folder URL of the trunk. This may sound wrong, but remember that the trunk is the start point to which you want to add the branch changes. You may also click to browse the repository.
In the To: field enter the full folder URL of the feature branch.
In both the From Revision field and the To Revision field, enter the last revision number at which the two trees were synchronized. If you are sure no-one else is making commits you can use the HEAD revision in both cases. If there is a chance that someone else may have made a commit since that synchronization, use the specific revision number to avoid losing more recent commits.
You can also use to select the revision.
This page of the wizard lets you specify advanced options, before starting the merge process. Most of the time you can just use the default settings.
You can specify the depth to use for the merge, i.e. how far down into your working copy the merge should go. The depth terms used are described in Section 4.3.1, “Checkout Depth”. The default depth is Working copy, which uses the existing depth setting, and is almost always what you want.
Most of the time you want merge to take account of the file's history, so that changes relative to a common ancestor are merged. Sometimes you may need to merge files which are perhaps related, but not in your repository. For example you may have imported versions 1 and 2 of a third party library into two separate directories. Although they are logically related, Subversion has no knowledge of this because it only sees the tarballs you imported. If you attempt to merge the difference between these two trees you would see a complete removal followed by a complete add. To make Subversion use only path-based differences rather than history-based differences, check the Ignore ancestry box. Read more about this topic in the Subversion book, Noticing or Ignoring Ancestry .
You can specify the way that line ending and whitespace changes are handled. These options are described in Section 4.10.2, “Line-end and Whitespace Options”. The default behaviour is to treat all whitespace and line-end differences as real changes to be merged.
The checkbox marked Force the merge is used to avoid a tree conflict where an incoming delete affects a file that is either modified locally or not versioned at all. If the file is deleted then there is no way to recover it, which is why that option is not checked by default.
If you are using merge tracking and you want to mark a revision as having been merged, without actually doing the merge here, check the Only record the merge checkbox. There are two possible reasons you might want to do this. It may be that the merge is too complicated for the merge algorithms, so you code the changes by hand, then mark the change as merged so that the merge tracking algorithm is aware of it. Or you might want to prevent a particular revision from being merged. Marking it as already merged will prevent the merge occurring with merge-tracking-aware clients.
Now everything is set up, all you have to do is click on the button. If you want to preview the results simulates the merge operation, but does not modify the working copy at all. It shows you a list of the files that will be changed by a real merge, and notes files where conflicts may occur. Because merge tracking makes the merge process a lot more complicated, there is no guaranteed way to find out in advance whether the merge will complete without conflicts, so files marked as conflicted in a test merge may in fact merge without any problem.
The merge progress dialog shows each stage of the merge, with the revision ranges involved. This may indicate one more revision than you were expecting. For example if you asked to merge revision 123 the progress dialog will report “ Merging revisions 122 through 123 ”. To understand this you need to remember that Merge is closely related to Diff. The merge process works by generating a list of differences between two points in the repository, and applying those differences to your working copy. The progress dialog is simply showing the start and end points for the diff.
The merge is now complete. It's a good idea to have a look at the merge and see if it's as expected. Merging is usually quite complicated. Conflicts often arise if the branch has drifted far from the trunk.
![[Tip]](tip.png)
|
Tip |
|---|---|
|
Whenever revisions are merged into a working copy, TortoiseSVN generates a log message from all the merged revisions. Those are then available from the button in the commit dialog. To customize that generated message, set the corresponding project properties on your working copy. See Section 4.17.3.10, “Merge log message templates” |
For Subversion clients and servers prior to 1.5, no merge information is stored and merged revisions have to be tracked manually. When you have tested the changes and come to commit this revision, your commit log message should always include the revision numbers which have been ported in the merge. If you want to apply another merge at a later time you will need to know what you have already merged, as you do not want to port a change more than once. For more information about this, refer to Best Practices for Merging in the Subversion book.
If your server and all clients are running Subversion 1.5 or higher, the merge tracking facility will record the revisions merged and avoid a revision being merged more than once. This makes your life much simpler as you can simply merge the entire revision range each time and know that only new revisions will actually be merged.
Branch management is important. If you want to keep this branch up to date with the trunk, you should be sure to merge often so that the branch and trunk do not drift too far apart. Of course, you should still avoid repeated merging of changes, as explained above.
|
|
Tip |
|---|---|
|
If you have just merged a feature branch back into the trunk, the trunk now contains all the new feature code, and the branch is obsolete. You can now delete it from the repository if required. |
|
|
Important |
|---|---|
|
Subversion can't merge a file with a folder and vice versa - only folders to folders and files to files. If you click on a file and open up the merge dialog, then you have to give a path to a file in that dialog. If you select a folder and bring up the dialog, then you must specify a folder URL for the merge. |
Subversion 1.5 introduced facilities for merge tracking. When you merge changes from one tree into another, the revision numbers merged are stored and this information can be used for several different purposes.
-
You can avoid the danger of merging the same revision twice (repeated merge problem). Once a revision is marked as having been merged, future merges which include that revision in the range will skip over it.
-
When you merge a branch back into trunk, the log dialog can show you the branch commits as part of the trunk log, giving better traceability of changes.
-
When you show the log dialog from within the merge dialog, revisions already merged are shown in grey.
-
When showing blame information for a file, you can choose to show the original author of merged revisions, rather than the person who did the merge.
-
You can mark revisions as do not merge by including them in the list of merged revisions without actually doing the merge.
Merge tracking information is stored in the svn:mergeinfo
property by the client when it performs a merge. When the merge is committed
the server stores that information in a database, and when you request merge,
log or blame information, the server can respond appropriately. For the system
to work properly you must ensure that the server, the repository and all clients
are upgraded.
Earlier clients will not store the svn:mergeinfo property
and earlier servers will not provide the information requested by new clients.
Find out more about merge tracking from Subversion's Merge tracking documentation .
Merging does not always go smoothly. Sometimes there is a conflict, and if you are merging multiple ranges, you generally want to resolve the conflict before merging of the next range starts. TortoiseSVN helps you through this process by showing the merge conflict callback dialog.
It is likely that some of the changes will have merged smoothly, while other local changes conflict with changes already committed to the repository. All changes which can be merged are merged. The Merge Conflict Callback dialog gives you three different ways of handling the lines which are in conflict.
-
If you are merging text files then these first two buttons allow you to merge non-conflicting lines as normal and always prefer one version where there are conflicts. Choosing Prefer local will select your local version in every conflict, i.e. it will prefer what was already there before the merge over the incoming change from the merge source. Likewise, Prefer repository will select the repository changes in every conflict, i.e. it will prefer the incoming changes from the merge source over what was already in your working copy. This sounds easy, but the conflicts often cover more lines than you think they will and you may get unexpected results.
If your merge includes binary files, merging of conflicts in those is not possible in a line-by-line mode. A conflict in a binary file always refers to the complete file. Use Prefer local to select the local version as it was in your working copy prior to the merge, or Prefer repository to select the incoming file from the merge source in the repository.
-
Normally you will want to look at the conflicts and resolve them yourself. In that case, choose the which will start up your merge tool. When you are satisfied with the result, click .
-
The last option is to postpone resolution and continue with merging. You can choose to do that for the current conflicted file, or for all files in the rest of the merge. However, if there are further changes in that file, it will not be possible to complete the merge.
If you do not want to use this interactive callback, there is a checkbox in the merge progress dialog Merge non-interactive. If this is set for a merge and the merge would result in a conflict, the file is marked as in conflict and the merge goes on. You will have to resolve the conflicts after the whole merge is finished. If it is not set, then before a file is marked as conflicted you get the chance to resolve the conflict during the merge. This has the advantage that if a file gets multiple merges (multiple revisions apply a change to that file), subsequent merges might succeed depending on which lines are affected. But of course you can't walk away to get a coffee while the merge is running ;)
When you develop a new feature on a separate branch it is a good
idea to work out a policy for re-integration when the feature is
complete. If other work is going on in trunk
at the same time you may find that the differences become significant
over time, and merging back becomes a nightmare.
If the feature is relatively simple and development will not take long then you can adopt a simple approach, which is to keep the branch entirely separate until the feature is complete, then merge the branch changes back into trunk. In the merge wizard this would be a simple Merge a range of revisions, with the revision range being the revision span of the branch.
If the feature is going to take longer and you need to account
for changes in trunk, then you need to keep
the branch synchronised. This simply means that periodically
you merge trunk changes into the branch, so that the branch
contains all the trunk changes plus the
new feature. The synchronisation process uses
Merge a range of revisions. When the feature
is complete then you can merge it back to trunk
using either Reintegrate a branch or
Merge two different trees.
Another (fast) way to merge all changes from trunk to the feature branch is to use the → from the extended context menu (hold down the Shift key while you right click on the file).
This dialog is very easy. All you have to do is set the options for
the merge, as described in Section 4.20.3, “Merge Options”.
The rest is done by TortoiseSVN automatically using merge tracking.